15th February 2010 at 8:40 am #5030ISMAEL YOUNIS ABUJARADMember
Dear friends , Ladies and gentlemen,
I am conducting a study in which I am looking into the relationship between Organizational Innovativenss as my IV and Organizational Performance. I measure Orgnizational Performance using Profitability, sales growth and employee turnover. Profitability was measured using BPCI , the average values of ROI + ROA + ROS / 3 (3 means last three years) ..So, my DV is not a scale. I got BPCI using excel then i pasted the values in SPSS. When I ran regression, My model was not significant. Any reason why? do i have to change the values of my DV into a scale too ? what to do ? I appreciate your input19th February 2010 at 4:17 pm #5032Dr. James LaniParticipant
Overall, the reason a multiple regression is not statistically significant is because the predictors do a poor job of explaining the variability in the DV. i don’t think the problem is the combining of the three scales to create your DV; are each of the IV’s averaged over the three years as well? we’d be glad to speak with you about multiple regression.
If you would like more information on regression, see my website at:
Dr. James Lani
Statistics Solutions19th February 2010 at 6:25 pm #5031ISMAEL YOUNIS ABUJARADMember
Dear Dr. James Lani,
first of all, let me put you into the picture of what i am doing. My IV is organizational innovativness. There are five dimensions that I am adopting from Wang and Ahmed (2004). They are Product innovativeness, process innovativeness, strategic innovativeness, market innovativeness, and behavior innovativeness. Well, i ran factor analysis and reliabitliy. there was no problem with that. My DV is financial performance. One dimension to measure it was profitability. I used Business Performance Composite Index (BPCI) where I asked the housing developers to provide me with the ratios in percentage about ROI, ROA, ROS for 2006,2007,2008 and then i summed up the average and divided by three. I pasted the ratios on SPSS and ran regression between teh IV s and the DV. When the sample size was 57, the model significance level was 0.58, i was happy and i said to myself when the sample size is bigger, then there is probability that the model becoems significant, but unfortunately it did not. I am disappointed very much disappointed indeed. I have to put the values of the DV into a scale , just try to see whether it will make any difference, but it did not make any difference.
- You must be logged in to reply to this topic.