quantitative analysis in sociology

Home Forums Default Forum quantitative analysis in sociology

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
    Posts
  • #3716

    I am doing PhD in Sociology . I have used statistical tests such as chi square to check the significance of variables and logistic regression to see the effect on a binary outcome Can some one read my results chapter and comment on the substance of analysis and the style to improve the presentation. Thanks. Regards my mail add is “mainpal_83@yahoo.co.in”. Looking forward to some signal of help.

    #3726
    Jeremy Miles
    Participant

    Don’t you have a supervisor for that?

    #3725

    I am doing my PhD as a part time student and the guide is not used to newer tools like SPSS. The request is to link with people who handle newer technology and can advise on validity, substance of statistical tests in a given setting.Your derogatory and demeaning comment has hurt me. I feel embarrassed to approach next time for help. Regards.God Bless you.

    #3724
    Jeremy Miles
    Participant

    Apologies for hurting you.  It was not intended to be demeaning or derogatory.  On occasions I have read students PhD chapters as a favor, the student then changes it based on my comments, and gives it to the supervisor.  The supervisor says “Why did you put this, I don’t like it.”  The student comes back to me and asks why I said what I said.  It all gets very confusing and the student gets stuck in the middle of an academic debate that they don’t understand. 

     

    However, if you still want to ask for help, instead of asking someone to read the whole thing, why not post some sections – a paragraph, say.  Put what you have put in the paragraph and ask if it’s phrased OK and applied appropriately.

     

    Jeremy

     

    P.S.  SPSS isn’t a newer tool – it’s been around since 1968.

     

     

    #3723

      Dear Prof Sending you my work for comments. You are reuested to advise improvements as you deem fit. Your comments will have lot of value to me. They shall also be confidential. I shall not revert back to you with my guide’s observations. Thanks and regrds. Chapter 5

    RESULTS AND ANALYSIS

    A perception is held by many including the defence services personnel that present day Gujarati youth is less inclined to join as soldier. The author has come across very few Gujarati soldiers in his career of 28 years. This study therefore was conceived to investigate if the said perception was a fact or myth and explore the propensity of Gujarati youth in recruitable age to volunteer for the Army by a field survey.
    The focus of this study was to estimate the present day propensity of Gujarati male youth to enlist themselves in the Indian Army as a sepoy/Jawan (PBOR) . The endeavour was to trace back history of Gujarati society for its participation in military duties or soldiering. The propensity for military participation in volunteer armies of modern times to some extent though not fully depend on historical tradition of regional/social class warriorship tendencies (cultural aspects), availability and contact with role models and the ultimate factors of individual choice and characteristics. Hence this study examined the history of valour / warriorship demonstrated by Gujarati castes/ communities in chapter on military history of Gujarat to trace military related tendencies as reported / observed by authors in the past.
    This chapter on results and analysis of the sample data deals with two broad statistical analyses; first the descriptive statistics about the profile of respondents and secondly the measure of propensity of Gujarati youth (boys) to join the army. The data was collected for school going boys and their parents on similar questionnaire suitably modified as per category of respondent (boy or a parent). A total of 179 school going boys and 169 parents have answered the questionnaires. Data from 348 respondents was realized against the planned figure of 400. Once the field work had been completed, the questionnaires were edited and coded. The coded data was filled in the computer software SPSS11.5 and searched for coding errors. Some categories were recoded where required.
    This chapter on results addresses the three questions of the research problem. The first aims to find the demographics (a range of lower and upper measures, overall response percentages to categories being measured) of different demographic variables of the group sample to denote the composition of the subgroups ; followed secondly by identification of the significant predictor variables that have a significant association with outcome variable and thirdly to see the effect of significant independent variables on the binary decision of joining (answer ‘yes’ or (1) to join as an army soldier and answer ‘no’ or (0) for not intending to join as a soldier). The first consideration of analysis will give the descriptive values of the sample on various social demographic and personal habits/ characteristics. The second and the third questions will provide measures of association (inferential statistical results) about significance and the effect size of independent variables (characteristics) by a regressive procedure. The data analysis has been done by using SPSS 11.5 computer software. The effect sizes have been measured as odds ratios. The missing/not answered responses are included in figures/charts and tables to indicate number of respondents who did not answer a particular question whereby being indicative of likely disinterest in the query or lack of knowledge.

    The young job seekers and their parents have been profiled demographically for their ethnicity (Gujarati or non Gujarati), district of residence, age, education, caste, preferences for various job types with a high school qualification, their daily life style routine (time devoted to physical activity, religious practices, leisure activities), visit to a military cantonment and academic performance, at individual level. The data on number of times a boy has made an effort (more number of visits to a recruiting office/attending recruitment rallies implying his greater interest) to get recruited or otherwise also indicates his liking/disliking for the army service. The parameters in this category are designed to seek an estimate of likely impact of individual actions/attainments/choices on a young man’s willingness/intention to join the army service. They represent the personality factors of individuals. David E Super in his book titled “The Psychology of Careers” (1957) has included ‘individual characteristics’ as one of the major factors responsible for determining the choice of careers.
    The second descriptive aspect of the sample’s group profile relates to the perceptions/beliefs/attitudes/knowledge of youth and their parents about the army, about the societal aspects in the form of volunteer castes for the military service in their neighbourhood, non volunteer castes, implications of becoming a soldier, appeals for joining the army, awareness and knowledge about the job content and conditions of service of the army soldier. The data on the type of communications referred by the high school youth and their parents in seeking information about the army recruitment have also been recorded. This data can be utilised for evolving best mix of communication media to reach the target population. This part of the results will show the interaction and communication between the prevalent socioeconomic environment and intention/knowledge/beliefs of the participants about making army a career.
    Influences and actions in the form of availability of a role model in the family, hearing of folklores depicting stories of war / war heroes/warriors (possible indicator of cultural tradition of military participation by Gujaratis), working women culture and maximum education in the household locality to indicate the effect of such economic and educational indicators in their immediate neighbourhood on the propensity of youth to join as soldiers. Prominence of certain social groups (castes) in the community is indicative of existing social hierarchy in villages and social mobility to nearby places in search of jobs indicates absorption of locals into the local economic activities around their places of residence. The details of these economic activities have been recorded as type of work/jobs available and the wages thereof. Family consumption and wealth levels have been measured in the form of land holdings, type of residential unit, that is, a built -up house or a makeshift arrangement, possession of mechanical farm equipment like a farm tractor, commodities of daily use like television, mobile phone, and a motorcycle. Availability of social infrastructures such as village schools, health centres, closeness to a highway, rail head (indicative of proximity to means of communication i.e. indicator of regional development), location in kilometers of an industrial unit, number of people employed in such units are indicators of a given region’s overall economic development status. It is assumed that economic well being of survey participants is likely to impact decision to join the army negatively.

    Part 1: Profile of Job Seekers and their Cultural and Social Indicators

    The sample profile data has been divided in three categories- Individual Parameters, Group Perceptions and Social and cultural Influences parameters for the descriptive analysis.
    Individual Characteristics of the Group

    A total of 348 respondents in the sample had 318 Gujarati or 91.4 percent and 30 non- Gujarati or 8.6 percent participants. The sampled group composed of 179 candidates (51.4 percent) and 169 parents (48.6). The district wise breakdown of participants was as per Table 5.1.
    Table5.1 District of Residence

    District Count Percent
    Baroda 46 13.2
    Surat 46 13.2
    Sabarkantha 42 12.1
    Godhra 40 11.5
    Gandhinagar 38 10.9
    Banaskantha 32 9.2
    Nadiad 44 12.6
    Ahmedabad 60 17.2
    Total 348 100.0

    Finding: The data suggests that the stakeholders are evenly distributed amongst young boys group and parents group thereby eliminating a bias in the information.
    Age Profile of Boys

    Age in number of years completed was recorded. 90.49 percent of the sample represented ages between 15 and 21 years. A combatant soldier can join the army up to the age of 21 years. 136 boys out of 179 were between 15 and 18 years of age, that is, about 75.97 percent. 26 boys or 14.52 percent were in the age bracket of 18 to 21 years of age. 11 participants or 6.14 percent of sample size had ages between 22 and 24 who are eligible to get recruited as clerks and tradesmen and likely to be undergraduate students or graduates

    Table 5.2 Age of respondents: Boys
    Age in Years Count Percent
    Not Answered 1 .6
    13 2 1.1
    14 3 1.7
    15 32 17.9
    16 40 22.3
    17 40 22.3
    18 24 13.4
    19 8 4.5
    20 8 4.5
    21 10 5.6
    22 7 3.9
    23 3 1.7
    24 1 .6
    Total 179 100

    Finding: The age profile of boys was representative of the age eligibility to get recruited in the army as soldier. 90 percent of the boys were in the age bracket of 15 to 21 years, which is the bracket for enrollment as combat soldiers.

    Age of Parents Age bracket 40 -50 years constituted 84.61 percent among parents. 14 parents constituting a small fraction of 8.28 percent were of ages between 51 and 60 years. Remainder parents that are a small minority with approximately seven percent were in the age bracket of 34 to 39 years.

    Educational Profile of Boys
    The boys and parents were instructed to give the number of years that they had taken formal education. Total 82.68 percent boys were from 10 to 12 years of formal education. 14.52 percent of sampled boys had 13 to 15 years of formal education thereby implying that they were post higher secondary students. See Table 5.3 for educational profile of boys.
    Table5.3 Education of Boys
    Education in Years Count Percent
    8 1 .6
    9 1 .6
    10 57 31.8
    11 55 30.7
    12 36 20.1
    13 15 8.4
    14 7 3.9
    15 4 2.2
    16 1 .6
    17 2 1.1
    Total 179 100

    Findings: Majority of the boys were from the classes set as criteria to be considered for recruitment in combat or non combat roles. ‘Educational conditions’ was identified by E Ginzberg in his seminal book titled “Theory of Occupational Choice” (1951) as a factor affecting occupational choice. It implied that persons with better educational attainments had bigger aspirations of a job.
    Education of Parents
    70.4 percent of 169 parents had 10 or less years of formal education. 29.5 per cent had education higher than 10 years of formal education. Majority of the parents were non-high school graduates or less educated.
    Findings: Less education of the parents is a driver to motivate parents to send boys to become soldiers. Report of Paul Sackett and Anne Mover of Committee on the Youth Population and Military Recruitment found that parents education attainment, increased youths resources and aspirations for education. Therefore it was expected that less educated parents would prefer their wards to get into a job rather than a college. Previous studies show that those with lower family incomes, larger family sizes (more sharing of scarce resources) and less educated parents are more likely to join the military (Asch et al, 1999; Kilburn & Asch, 2003, Kilburn & Klerman, 1999). Ronald Inglehart’s study in 1976 revealed that better educated parents were less inclined about a son’s military service.

    Table 5.30 Knowle Table 5.30 Knowledge about salary of a Jawan
    Salary Per Month Frequency Percent
    More thanRs15000 110 31.6
    Less than Rs 5000 105 30.2
    Rs 5000 6 1.7
    Rs 6000 13 3.7
    Rs 7000 10 2.9
    Rs 8000 11 3.2
    Rs 9000 15 4.3
    Rs 10000 5 1.4
    Rs 12000 48 13.8
    Rs 13000 1 .3
    Rs 15000 24 6.9
    Total 348 100

    Job Content in the Army Half the respondents either did not know or did not answer the question. Five categories of nature of jobs namely border/border duty; combat/fight enemy/kill enemy; defend the nation; physical work; and hard work constituted a combined 37.5 percent of the responses. These two broad category groups constituted 88.4 percent. Remainder about 12 percent respondents gave different versions of work content as perceived by them. A study by Gerhart (1991) indicated that the ‘features of the perspective job’ impacted the decision to join an organisation.
    Findings: The dismal performance of respondents as regards knowledge about terms and conditions of army service indicates either lack of interest in the service or advertent avoidance of army as a service career.

    Any other 2 .6 92.2
    Rana Pratap’s story 14 4.0 96.3
    Prithviraj Chauhan’s story 8 2.3 98.6
    religious and clan tales 1 .3 98.9
    Shri Mahakali 1 .3 99.1
    Shivaji’s story 1 .3 99.4
    Subhash Chandra Bose’s story 1 .3 99.7
    Shahid Bhagat Singh’s story 1 .3 100
    Total 348 100

    Finding: The boys as well as parents did not mention a clan/community/regional story or a song describing a war or warriors.

    Part 2: Identification of Significance of Associations

    Part 1 of the results dealt with the descriptive analysis of the sample. It gave out the broad qualitative characteristics of the participants. Part 2 and 3 are inferential in nature and involve conducting statistical tests to seek relevant inferences about the significance of variables impacting the outcome and propensity of respondents to join the Army.

    Part 2 addressed the second research question. “What are the variables (characteristics/factors) which significantly impact or do not impact the propensity of Gujarati youth to join or refrain from joining the Army?”

    Chi square tests were conducted between deterministic variables and the outcome variable of the study-whether respondents wanted to join as soldiers or the parents were willing to send their son into the army service. The response to outcome variable was binary that is response 1 indicated ‘yes’ and response 0 indicated ‘no’. The significance levels thus identified by this bivariate test helped in including the significant independent variables in the logistic regression to measure the impact on outcome variable.

    Table 5.39 Significant Associations between Outcome and Variables

    Ser Variable Categories Chi square value df Significance 2 sided
    1 Candidate or father 2 39.47 1 .000**
    2 District of Residence 8 25.18 7 .001**
    3 Gujarati vs Non Gujarati 2 17.81 1 .000**
    4 Caste 6 14.45 5 .013**
    5 Caste Category 4 3.032 3 .387
    6 Age 37 71.88 36 .000**
    7 Education in years 18 46.79 17 .000**
    8 Occupation of father 4 8.116 3 .044**
    9 Income of father 4 8.88 3 .031**
    10 Academic Performance in 10th standard 5 21.87 4 .000**
    11 Academic Performance in 12th standard 5 18.83 4 .001**
    12 First Choice 14 24.29 13 .029**
    13 Army as first choice 13 31.02 12 .002**
    14 Exercise 2 1.85 1 .173
    15 Daily Prayers 4 4.50 3 .212
    16 Sportsman in the village 3 1.15 2 .562
    17 Shikar 3 2.13 2 .343
    18 School in the village 3 9.30 2 .010**
    19 School village upto which class 6 5.80 5 .326
    20 Relative in armed forces 3 3.63 2 .163
    21 Persons in service in neighbourhood 5 14.12 4 .007**
    22 Maximum Education in Neighbourhood 6 6.27 5 .281
    23 Visit to a camp 3 2.24 2 .325
    24 War Folklore 15 26.12 14 .025**
    25 Pensioner in the family 2 9.66 1 .002**
    26 Size of the family 12 50.72 11 .000**
    27 No of Earning members 6 19.57 5 .002**
    28 Landownership 3 10.99 2 .004**
    29 Type of house 4 28.92 3 .000**
    30 Tube well in Household 3 2.64 2 .267
    31 Television set 3 2.64 2 .267
    32 Motorcycle 3 .169 2 .919
    33 Mobile in the family 3 2.55 2 .279
    34 Industry near your village 3 28.01 2 .000**
    35 Distance of nearest Highway 6 24.29 5 .000**
    36 Part time job 3 2.39 2 .303
    37 Part of NCC 7 15.57 6 .204
    38 Patriotism as the appeal 10 25.52 9 .002**
    39 Salary of a soldier 11 24.87 10 .006**
    40 Why you want to join 8 16.43 7 .021**
    41 Started Job Search after which class 6 10.18 5 .070
    42 How long ago decision taken to join 4 24.54 3 .000**

    ** indicates highly significant association
    The deterministic variables that were found to have weak association were caste category, habit of daily exercise, daily prayers, sportsman in the village, trend for shikar or outdoor activities, school village upto which class, relative in the armed forces, maximum education in the neighbourhood, visit to a military camp, household consumption items like, tube well, television, motorcycle and mobile, doing part time job, being part of NCC/NSS and the class after which a boy started job search. However the results were in contrast to assumption made in the beginning of the study that these variables will strongly impact the decision of respondents and therefore included in the questionnaire.
    After the regression test some of the significantly associated independent variables with the outcome variable were excluded because of greater effect of other variables in the model. The ultimate model had just 17 relevant independent variables which had a significant impact on the outcome variable. The detailed results of regression test are covered in Part 3 of the chapter.

    Part3: Measurement of Propensity of Sample Group
    The research question answered in this part was “What is the direction of effect of significant variables to impact the propensity of sample respondents?”

    Test Diagnostics
    1. To prevent problems of non-convergence, prior to conducting the logistic regression procedure, frequencies for each of the variables were examined. According to Tabachnick and Fidel (2007), the non-convergence occurs when the combination of variables produces cells with zero cases. The following variables had categories with very few cases:
    a. Father’s occupation
    b. First choice of job
    c. School
    d. Folklore
    e. Distance from highway
    f. Academic achievement at 10th standard
    g. Academic achievement at 12th standard
    h. Part-time work experience
    i. Appeal of army service
    j. When decision to join army was made
    2. Thus, categories within the ten categories were collapsed so that each of the ten variables had fewer categories (and more cases per category).
    3. A stepwise logistic regression procedure (using the backward elimination likelihood ratio option) was conducted to determine which of the 36 variables predicted the likelihood of Gujarati boys joining the army. Indicator variables were created for variables that were measured on a nominal and ordinal scale. The reference category for each set of indicator variables was the first category (i.e., usually the “No Answer or Don’t Know” category).
    4. The stepwise procedure stopped at the 19th step (since removal of the least significant variable resulted in the model generated during the 18th step). The findings of the 19th step are shown in Table 5.40.

    Results
    5. The findings in Table 5.40 reveal inferences as given in the subsequent paragraphs.
    6. Education significantly predicted the propensity to join the army (Wald (1) = 7.44, p = .006). The negative coefficient suggests that for every unit increase in number of years spent studying, the propensity to join the army decreased by .82.
    Finding: The tendency of declining interest with the increase in academic standard is also reported by the authors of a technical report prepared in 2000 for US students of class 11th and 12 th which found that “there are at least some young men who look favourably on military service when they are sophomores (standard 11th) but whose views switch by the time they are seniors (standard 12 th)”
    7. Family size significantly predicted the propensity to join the army (Wald (1) = 17.29, p = .000). The negative coefficient suggests that for every unit increase in family size, the propensity to join the army decreased by .87.
    Finding: This is in contrast to several studies which found that having more siblings increased the likelihood of serving in the military.
    8. Folklore also significantly predicted the propensity to join the army (Wald (3) = 9.35, p = .025). However, none of the indicator variables testing differences between types of folklore heard vs. no folklore heard significantly predicted the propensity to join the army.
    Finding:
    9. Academic achievement at the 10th standard did not significantly predict the propensity to join the army. But in comparison to respondents who did not answer this question (or respondents who were still studying in the 10th standard), the propensity to join the army increased by 1.73 for respondents who were in the Third Division (Wald (1) = 5.67, p = .017).
    Finding: Army was considered a likely choice by those who were assessed weak in studies. The same sentiment was given by the respondents while answering a question on who joins the army (the answer being those who are weak in studies).
    10. Academic achievement at the 12th standard significantly predicted the propensity to join the army (Wald (1) = 6.28, p = .043). However, none of the indicator variables testing differences between type of achievement vs. no answer (or respondent/son still studying the 12th standard) significantly predicted the propensity to join the army.
    Finding: Participants with better performance in 12 th standard (first division) were negatively inclined to join as soldiers than the third divisioners who had a 8.7 times more inclination to join the army.
    11. The time when the decision was made significantly predicted the propensity to join the army (Wald (4) = 16.46, p = .002). First, in comparison to those who did not join (or who did not answer the question or who had not yet taken the decision), the propensity to join the army increased by 5.25 for those who decided to join the army three to six months before (Wald (1) = 8.95, p = .003). Second, in comparison to those who did not join (or who did not answer the question or who was not yet taken), the propensity to join the army increased by 5.83 for those who decided to join the army one year before (Wald (1) = 9.89, p = .002). Third, in comparison to those who did not join (or who did not answer the question or who was not yet taken), the propensity to join the army increased by 10.85 for those who decided to join the army three years before (Wald (1) = 11.28, p = .001).
    Finding: This is in agreement with the finding of the Committee on Youth Population and Military Recruitment in USA which reported that youth “make decisions well into their twenties”.
    12. First choice of a job significantly predicted the propensity to join the army (Wald (3) = 16.40 p = .001). In comparison to those who did not answer the question, the propensity to join the army increased by 5.73 for those respondents whose first job choice was to be a security guard, army Jawan, salesman, BSF Jawan, or some other job (Wald (1) = 6.51, p = .011).
    Finding: Respondents whose first choice of a job was a primary teachership had a negative propensity for the Army. However that was not significant statistically. Respondents from the category of security guard, army Jawan, salesman, BSF Jawan and any other job as first choice of a job were significantly inclined to join the army service.
    13. Distance from the nearest highway significantly predicted the propensity to join the army (Wald (3) = 8.74, p = .033). In comparison to those who did not answer the question, the propensity to join the army decreased by .09 for those respondents who lived 11 or more kilometers from the nearest highway (Wald (4) = 5.69, p = .017).
    Finding: This finding may be an outcome of being remotely located and not aware of army recruitment programmes in comparison to those who had better access to surface communication.
    14. Reason for joining the army significantly predicted the propensity to join the army (Wald (2) = 9.65, p = .008). First, in comparison to those who did not answer the question, the propensity to join the army increased by 10.10 for respondents who indicated patriotism as the primary reason for joining the army (Wald (1) = 8.86, p = .002). Second, in comparison to those who did not answer the question, the propensity to join the army increased by 6.10 for respondents who indicated some other reason as the primary reason for joining the army (Wald (1) = 7.52, p = .006).
    Finding: Both these categories describe the professionalistic attitude of respondents. That means they were guided by higher values than occupational or materialistic orientation towards the Army.
    15. Domicile significantly predicted the propensity to join the army (Wald (1) = 8.18, p = .003). In comparison to respondents who were not Gujarati, the propensity to join the army decreased by .08 for the respondents who were Gujarati.
    Finding: The army service was less popular among Gujarati youth and parents than respondents who were non Gujaratis.
    16. Exercise also significantly predicted the propensity to join the army (Wald (1) = 4.41, p = .036). In comparison to respondents who did not exercise, the propensity to join the army decreased by 2.32 for respondents who exercises.
    Finding:
    17. Whether or not a sportsman lived in the village significantly predicted the propensity to join the army (Wald (2) = 7.92, p = .019). In comparison to respondents who did not answer the question, the propensity to join the army increased by 12.56 for respondents who lived in a village without a sportsman (Wald (1) = 4.14, p = .042).
    Finding: It indicates that most of the respondents who did not have sportsmen as their role models chose the army service.
    18. Type of house did not significantly predict the propensity to join the army. But in comparison to respondents who did not answer the question, the propensity to join the army increased for respondents who lived in a kuchcha (Wald (1) = 5.52, p = .019).
    Finding: Economically weaker were more inclined to join the army service in comparison to people who had better house (pukka).
    19. Motorcycle ownership significantly predicted the propensity to join the army (Wald (2) = 7.10, p = .029). None of the comparisons, however, significantly predicted the propensity to join the army.

    20. Television ownership significantly predicted the propensity to join the army (Wald (1) = 6.78, p = .009). In comparison to respondents who did not own a television, the propensity to join the army increased by 4.72 for respondents who owned a television.
    Finding: This could indicate the increased awareness of youth and their parents if they owned Television sets in the house.
    21. Whether or not the respondent lived in a village that was located close to an industrial area significantly predicted the propensity to join the army (Wald (2) = 17.46, p = .000). First, in comparison to respondents who did not answer the question, the propensity to join the army decreased for respondents who lived in a village that was not close to an industrial region (Wald (1) = 16.19, p = .000). Second, in comparison to respondents who did not answer the question, the propensity to join the army decreased for respondents who lived in a village that was close to an industrial region (Wald (1) = 8.45, p = .004).
    Finding: The effect of closeness to an industry was neutral. Whether the industry was there or not Gujarati boys were significantly disinclined to join the army.

    Table 5.40
    Stepwise (Backward Elimination) Logistic Regression Results for Propensity of Gujarati Boys to Join the Army (N = 348)
    Variable B SE Wald df Sig. OR
    Education
    Family size
    Folklore
    NA/DK vs. Geeta and/or Ramayan
    NA/DK vs. Ramayan and/or Mahabharat
    NA/DK vs. all others
    Academic achievement at 10th standard
    NA/Still studying 10th vs. third division
    NA/Still studying 10th vs. second division
    NA/Still studying 10th vs. first division
    Academic achievement at 12th standard
    NA/Still studying 12th vs. third division
    NA/Still studying 12th vs. first division -.20
    -.13

    -1.02
    -.56
    1.30

    2.46
    .70
    1.43

    2.06
    -.90 .07
    .03

    .61
    .46
    .73

    1.03
    .90
    .87

    1.32
    .58 7.44
    17.29
    9.35
    2.84
    1.44
    3.19
    7.02
    5.67
    .60
    2.70
    6.28
    2.69
    2.38 1
    1
    3
    1
    1
    1
    3
    1
    1
    1
    2
    1
    1 .006
    .000
    .025
    .092
    .230
    .074
    .071
    .017
    .439
    .100
    .043
    .101
    .123 .82
    .87

    .36
    .57
    3.67

    1.73
    2.00
    4.20

    8.70
    .41
    Note. OR = Odds ratio. NA = Not answered. DK = Don’t know.

    Variable B SE Wald df Sig. OR
    When decision was made
    NA/Not taken/Not join vs. 3 to 6 months
    NA/Not taken/Not join vs. one year
    NA/Not taken/Not join vs. two years
    NA/Not taken/Not join vs. three years
    First choice for job
    NA vs. primary teacher
    NA vs. constable, CRPF Jawan, clerk, bus driver/conductor, fire brigade, factory worker, peon
    NA vs. security guard, army Jawan, salesman, BSF Jawan, other
    1.66
    1.76
    2.38
    1.05

    -.16
    ..58

    1.75
    .55
    .56
    .71
    .60

    .68
    .71

    .68 16.46
    8.95
    9.89
    11.28
    3.01
    16.40
    .06
    .65

    6.51 4
    1
    1
    1
    1
    3
    1
    1

    1 .002
    .003
    .002
    .001
    .083
    .001
    .815
    .419

    .011
    5.25
    5.83
    10.85
    2.85

    .85
    1.78

    5.73
    Note. OR = Odds ratio. NA = Not answered. DK = Don’t know.

    Variable
    B SE Wald df Sig. OR
    Distance from nearest highway
    NA vs. 1 to 5 km.
    NA vs. 6 to 10 km.
    NA vs. 11 or more km.
    Reason for joining army
    NA vs. patriotism
    NA vs. all other reasons
    Non-Gujarati vs. Gujarati
    No exercise vs. exercise
    Sportsman in village
    NA vs. no
    NA vs. yes
    .02
    -.90
    -2.37

    2.31
    1.81
    -2.53
    .84

    2.53
    1.62
    .84
    .99
    .99

    .78
    .66
    .89
    .40

    1.24
    1.26 8.74
    .00
    .84
    5.69
    9.65
    8.86
    7.52
    8.18
    4.41
    7.92
    4.14
    1.64 3
    1
    1
    1
    2
    1
    1
    1
    1
    2
    1
    1 .033
    .977
    .356
    .017
    .008
    .003
    .006
    .004
    .036
    .019
    .042
    .200 .
    1.02
    .41
    .09

    10.10
    6.10
    .08
    2.32

    12.56
    5.05
    Note. OR = Odds ratio. NA = Not answered. DK = Don’t know.

    Variable
    B SE Wald df Sig. OR
    Residence
    Baroda vs. Surat
    Baroda vs. Sabarkantha
    Baroda vs. Godhra
    Baroda vs. Gandhinagar
    Baroda vs. Banaskantha
    Baroda vs. Kheda
    Baroda vs. Ahmedabad
    Type of house
    NA vs. jhopda
    NA vs. kuchcha
    NA vs. pukka
    Family own motorcycle
    NA vs. no
    NA vs. yes
    -1.29
    -.22
    -1.01
    .90
    .24
    1.10
    .35

    -.19
    1.88
    .36

    3.08
    2.22
    .89
    .99
    .93
    .93
    .94
    .90
    1.02

    3.11
    .80
    .57

    1.62
    1.65 12.79
    2.10
    .05
    1.18
    .94
    .06
    1.25
    .11
    5.82
    .00
    5.52
    .41
    7.10
    3.64
    1.82 7
    1
    1
    1
    1
    1
    1
    1
    3
    1
    1
    1
    2
    1
    1 .077
    .147
    .825
    .277
    .332
    .801
    .263
    .735
    .121
    .950
    .019
    .520
    .029
    .056
    .177
    .28
    .80
    .37
    2.46
    1.27
    2.74
    1.41

    .82
    6.56
    1.44

    21.79
    9.21
    Note. OR = Odds ratio. NA = Not answered. DK = Don’t know.
    Variable
    B SE Wald df Sig. OR
    Family doesn’t own TV vs. family owns TV
    Industry near village
    NA vs. no
    NA vs. yes 1.55

    -6.62
    -4.89 .60

    1.65
    1.68 6.78
    17.46
    16.19
    8.45 1
    2
    1
    1 .009
    .000
    .000
    .004 4.72

    .00
    .01
    Overall model χ2 (41) = 221.99, p < .001
    Note. OR = Odds ratio. NA = Not answered. DK = Don’t know.

    #3722
    Jeremy Miles
    Participant

     

    —-begin section—

     

    Education significantly predicted the propensity to join the army (Wald (1) = 7.44, p = .006). The negative coefficient suggests that for every unit increase in number of years spent studying, the propensity to join the army decreased by .82. 
    Finding: The tendency of declining interest with the increase in academic standard is also reported by the authors of a technical report prepared in 2000 for US students of class 11th and 12 th which found that “there are at least some young men who look favourably on military service when they are sophomores (standard 11th) but whose views switch by the time they are seniors (standard 12 th)”

     

    —end section—

     

    You did a stepwise regression.  This is possibly problematic.  Search the internet for ‘problems with stepwise regression’.

     

    In the text you report the Wald test, which is not interesting, but you don’t report the effect, which is interesting – the reader is forced to go to the table to see the size of the effect.

     

    You need to be very careful not to assign causal statements in your writing. You are implying that increasing academics leads to a reduced propensity to join the army. You don’t know that that is the case, you only know that those with higher academics are less likely to join, you don’t know why.

     

    You discuss your findings in relation to another report.  I would save that for the discussion.  In the results, you just discuss your results. In the discussion you can talk about what they mean, in relation to other reports.  If you want to speculate about a possible causal relationship, then the discussion is the place to do it.

     

     

     

     

     

     

     

     

    #3721

    Dear Prof Miles thank you for your posts. I have taken note of your observations. Just some basic questions. Where should I write my discussion- as part of Results and Analysis chapter or include in the Conclusion chapter? My results chapter is about 56 pages. Should I remove some of my descriptive analysis text? where and how do I merge the analysis derived in demographic part and the statistical test part? Thanks and regards. col m p singh

    #3720
    Jeremy Miles
    Participant

    Usually you have a results chapter, where you present what you found, and then a discussion chapter where you discussion the implications of what you found.  Sometimes it’s worth combining results and discussion, and you need to make that explicit, but then you still need a discussion where you can discuss, for example, the limitations of your work.

    #3719

    Thanks Prof miles. You have been very prompt.Regards

    #3718
    Muir Houston
    Member

    You still have a major problem by using a stpewise entry method for variable selection – you would not get past any supervisors i know with such an approach – and then it is even less likely that you will convince an external examiner of the use of such a ‘fishing expedition’

     

    have you done as suggested by Jeremey in his useful responses to you and googled ‘problems with stepwise regression’?

    #3717

    Dear Prof  could you please have a look at the conceptualisation of my empirical study. I have attached the two tables which are part of my Review of literature and chapter on methodology respectively. I have redone my quantitative work as per your advice by enter method of logistic regression. I shall be very grateful if you could spare some time to give your feedback on these two tables. Thanks and regards. Col M P Singh

Viewing 11 posts - 1 through 11 (of 11 total)
  • The forum ‘Default Forum’ is closed to new topics and replies.