Hi Dear researchers
I Ask you to help me to solve a problem found in my statistic, the information will be shown bellow:
My thesis titled: "studying the impact of manager's competency on organizational loyalty, a case study"
in my survey: I have first two variables with each one items, the number one is independent variable called "manager's competency" with three sub-variables, second: variable called "organizational loyalty", the distribution of two variables are not normal distributed
what is the alternative of "one sample t test" if variables are non normal distributed? if I want to study the existence of each variable before studying regression.
You knew the one sample t test require: normal distribution, sample at least 30 etc...
Hi Abdelkarim Ziane
The one-sample Wilcoxon signed rank test is a non-parametric alternative to one-sample t-test when the data cannot be assumed to be normally distributed. It’s used to determine whether the median of the sample is equal to a known standard value (i.e. theoretical value). You can easily do that in R (link below).
http://www.sthda.com/english/wiki/one-sample-wilcoxon-signed-rank-test-in-r
Thank a lot
can I do this test with spss? because I am using this tool, but R tool I dont know how to use it
Hi Abdelkarim Ziane
The link below will run you through how to do it in SPSS. Best wishes
https://m.youtube.com/watch?feature=youtu.be&v=qTmfCjcRKls
thanks alot
But I have computed my variable called "organizational loyalty" from his items (items are questions of survey to employees) with spss, when I have got my variable I have tested him with explore to know if he is normal ditributed or not, the result was that he is not normal distributed, then I can't use one sample t test because of requiring normal distribution of variable which why I was asking for alternative
its seems that explanation of your test videos shows that they use two variables, for me I have just one variable
Hi Abdelkarim Ziane
As previously stated, the one-sample Wilcoxon signed rank test used to determine whether the median of the sample (in this case, your sample- organizational loyalty. Note that , the variable used in the video is: How Scientific accounting) is equal to a known standard value (i.e. theoretical value). Thus, you need to have a hypothetical/assumed median which you want to compare your variable (organizational loyalty) with. Best,
The video used only one variable (how scientific accounting) and assumed a reference/standard value of 2.5 for comparison.
Hi Eddie-Williams Owiredu
I have computed the median of organizational loyalty which is 3,33 , I have chose in spss:
Analyse+non parametric tests+one sample after that I have select in window:
1 in objective tab I have chosen custom analysis
2 in Fields tab I have selected my variable "organizational loyalty" from filed to test fields
3 in settings tab I have selected compare median to hypothesis (Wilcoxon signed-rank-test) and add 3,33
after run button I got this:
(additional information, all my items of organizational loyalty was with likers scale from 1 totally not agree 2 not agree 3 incertain 4 agree 5 totaly agree)
do I have chosen the correct test?
if yes what is the best interpretation? is employees are loyal? or disloyal?
If I am correct they are not
Hi Abdelkarim Ziane
If you have chosen a median value of 3.33 (which lies around (3)-uncertain-meaning that you assume that workers are uncertain about the organizational loyalty), then you did it right. The null hypothesis (Ho) for the test is that your chosen standard value (here, 3.33) is the same as the median for your collected data (organizational loyalty). So since the p-value is 0.660, you fail to reject the Ho. Hence, there is no statistically significant difference between the hypothesized/standard median (3.33) and the obtained median (from your study).
NB: You can double-click on the SPSS result to get the model viewer in order to see that actual median of your organizational loyalty data.
For the interpretation: Given an assumption that workers are uncertain about organizational loyalty, your data indicate neither agreement nor disagreement. The workers (from your data) are also uncertain about organizational loyalty.
Best,
Hi again Eddie-Williams Owiredu
Does I must chose an other median like 2? to hypothesis H° as no loyalty observed in this company.
when I chose the test it will be two values, one chosen by me, other calculated with spss, I will send you again all my results both
Please I have send you before same results, its supposed to be one result second with viewer I will generate them with PDF file, ok?!!! It was a mistake.
I have attached two files, please check, help me to understand and for interpretation
Hi Abdelkarim Ziane
If you choose a reference median value of 2, you assume that workers are in disagreement with the organizational loyalty.
The p-value now is 0.000, which indicate that there is a statistically significant difference between the hypothesized/standard median (2) and the obtained median (3.33).
For the interpretation: Given an assumption that workers are in disagreement with organizational loyalty, your data indicate that the workers (from your study) are still uncertain about organizational loyalty (since the median value was 3.33 which is closer to 3-uncertain than 4-agree).
Best,
Thanks a lot Dear Eddie-Williams Owiredu you have save me, I was worry and tired for searching solution, Thanks also to researchgate
I am at your disposal for any help especially in the field of computer, word, excel, maintenance pc (soft or hard)
wish you best, Best regard Abdelkarim Ziane
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression