How to Do a One-Tailed, One Sample t Tests in SPSS (10-5)
Вставка
- Опубліковано 15 січ 2025
- Conducting a one-tailed, one sample t test in SPSS is possible, if you know what parts of the output that you can and cannot use. This example uses a story about Ritalin to look for changes in attention in only one direction, using a one-tailed test. This example uses fictitious data. The example references an effect size calculator that can be used to determine critical values and precise p-values for both one- and two-tailed tests at a variety of alpha levels.
Effect size calculator for t Tests: drive.google.c...
This video teaches the following concepts and techniques:
How to run a one-sample t test in SPSS using the one-tail option
IBM SPSS Statistics software
Link to a Google Drive folder with all of the files that I use in the videos including the Effect Size Calculator for t Tests and datasets. As I add new files, they will appear here, as well.
drive.google.c...
Thank you Dr. Todd Daniel for the excellent teaching, and your material and way of teaching is always enlightening.
Thank you once again.
This video was so helpful for me. Thank you so much for posting this!
Best explanation video I've seen so far! Thank you!
You're very welcome!
Dedicating my thesis to you, pinky promise. Thanks dude!
Rock on! Please do and send me a copy...I'd love to see that!
Your videos are so explicit and very convenient to learn from.
I have particular concerns; how and where can I derive the population mean with which to compare my sample mean, in my study?
Thanks alot man. Saved my life
Excellent explanation....thank you
Glad it was helpful!
Hi! Is it correct to assume the reason why you used a 90% CI value on SPSS for the "One Sample T-Test" even though the given alpha is set at 0.05 is because it is a one-tailed test?
Since we only need to disprove
μ ≤ 105, (meaning we only care if the results are only greater than 105), we only need to consider one side of the normal distribution curve/one side of the fence (i.e. the right side). Hence, we divide the supposed "10%" alpha into 5% for each "Outside the Fence". Is my assumption correct?
P.s. Thanks in advance! Been binge-watching your series as part of my New Year's Resolution!
Love the binge watching...yes, the 90% is a workaround to force 5% into each tail. You ignore the lower tail and focus on the one tail you are interested in. That one tail holds 5%. It is the only way to get the info because SPSS does not have a specific one-tail setting.
Thanks for this video!
My pleasure! Thank you for commenting
how about two sample t test? can we use the similar way?
Hi there, it was really helpful for me. I have a query. What if my alternative hypothesis is μ < 3. Do I have to follow a different procedure?
If your alt is μ < 3, then you are doing a one-tailed test. It is distinctly possible that you will be doing a different type of test (independent sample t test, ANOVA, etc.) but the principles of interpretation will be the same (i.e. if the mean changes to any value greater than 3, it is non-significant regardless of how large it goes. Only values less than 3 will be potentially significant). Good luck .
For 90% confidence interval , level of significance is 10%. ie 0.1 for one tail but for two tail 0.05. In your tutorial the significance level for two tail shown 0.083, please let me clear.
Can we use this method for one-tailed Paired Samples t-Test?.
Yes, the concepts also work for paired samples. Be sure, however, that it makes sense to look for changes in only one direction, such as comparing before & after scores where it only works if the ending scores go up, for example
Many of the sample files you are using do not appear in the Google Drive that is currently available as of 12-28-19
You are absolutely right. I added the missing files to the folder. Look again and see if you find the file you need. Thank you for letting me know.
What if the resulted t - value was negative ? Can we take the absolute value and then compare it with the critical values?
A negative t value is interpreted the same way as a positive t value. The negative sign just means that your first group mean was smaller than the second group mean. Personally, I always ignore the sign and just report a positive t value. So, yes, take the absolute value (i.e. positive) and compare it with the C.V.
For downloading Rstats Effect Size calculator for T tests, pls give me the download link
Check the description for the video. I have a link to a Google Drive folder. You can right-click to download it. No need to ask for edit permission, just download directly.
Hi! thank you so much for this video! I just have a question if my p value in spss (2 tailed) shows a value of .000 will I use that to divide it by 2 or use .001 instead to divide it to 2?
If you hover your cursor over the .000 in SPSS, it will display the full p-value (i.e .0002618)...that might help. Or just go with .001 and call it a day. :o)
@@ResearchByDesign thank you so much!!
How would one explain the procedure when used in a publication?
Thanks for the question...I would need to know more about your specific project to answer. You probably saw the APA Style write up for this one sample t test at 15:39. I would include some information about the data cleaning that you did, assumptions checks, etc. but the write up would follow pretty closely to the example in the video. I wish you the best on your publication.
What if the same size is greater than 30 and we don't have population standard deviation and one tailed
If you sample size is greater than 30 and/or you do not know sigma (pop. SD), you can still use a one sample t test. You can also choose to conduct the t test as either one tail or two tail.
In the final write up, shouldn't the confidence interval be 90% instead of 95%? But, anyway, great video!
You would think so, but the 90% CI, put 5% in each tail. With a one-tailed test, we are looking at the 5% in only one tail. Using 90% CI is a work-around to get the 5% one-tailed results. Therefore, it is the same as a one-tailed 95% CI.
@@ResearchByDesign so, 98%CI for two sided test means 2.5% in each tail?
@@ozgurackaln4314 I think it would mean 1% in each tail.
@@ResearchByDesign Thank you for your easily understood videos. However can you please answer a follow up to the question above. In presenting your result in a table form, with many variables, do you use the lower and upper confidence bounds for 90% confidence interval or would you have to use the bounds for a 95% confidence interval.
I see from the APA style write up that you repeat the confidence bounds for 90% confidence Intervals but relabel it as 95% Confidence Interval. Is that right?
@@opisthognathous Think of this technique as a workaround to get the data you really want. While the 95% CI splits the 5% into 2.5% in each tail, using 90% CI, you split the 10% into 5% for each tail. You could report the lower CI and the upper as infinity; as long as the lower CI is not negative (or on the opposite side of zero) you would interpret the findings as statistically significant.
u coulda explained this in like 30 seconds
بننیزتنیحیکیکز پ بدبحیجچقی
How about two sample t test? can we use the similar way? What is to be done then?
I discuss that a little in my video about independent samples t test: ua-cam.com/video/6eqm7zdPc3U/v-deo.html. Look for the section about interpretation. You see the critical value for a one-tail test in the video (or on your own t table), you interpret the findings only if the direction of change is the same direction you predicted with your one tail. Good luck!