One-sample t-test - SPSS (Part2)

(Part 2) I demonstrate how to perform a one sample-test in SPSS using two different examples. The one sample t-test is appropriate for testing hypotheses about the difference between a sample mean and a population "mean" (or "mu") when the population standard deviation (sigma) is not known.

[@nitinsabharwal9177]

thankyou!!! well explained

[@ANWAR1963]

thank you well said

[@kahemchu3505]

the p-value for 2-tailed, do we need to divide it by two before we compare to 0.05? for example, if the p-value from table is 0.568, but its for 2 tailed, do we need divide it by 2, 0.568/2=0.284 to compare to 0.05

[@user-cy7eg3gt4x]

I m comparing my values of data but did't find it significant..but in literature the two variables have relationship..so what should i do?

[@how2stats]

Estimate statistical power. It's possible that you're sample size was not sufficiently large to have a good chance of finding a statistically significant effect.

[@richardsanchez4657]

this video has one flaw: There is a mismatch between the p-value examined and the sign in Ha. This researcher looks at the two-tailed p value, then Ha should NOT be directional. He should have changed Ha from "...greater than 100" to "...different than 100." If he wants to stick to the directional Ha "...greater than 100," he should look at 0.057/2= 0.0285 as the p value. This is a minor point that many graduate students struggle with. So, do not worry too much about it.

[@dralifenwick6993]

True.I noticed this as well.

[@roitorcino7076]

You used data from a Likert scale, which is an ordinal data. Why didn't you opt to use one sample k-s test?

[@how2stats]

As I review in my book (how2statsbook.com; Chapter 6), most parametric statistical analyses typically estimate fairly accurate standard errors and p-values when the ordinal data have at least five points.

[@roitorcino7076]

@@how2stats ohhhh thank you for a fast response. One final question, what statistical tool should I use if I'm finding relationship between nominal data (public/private/semi-private) and ordinal data (10k-20k, 21-k-30k)? I want to know if there's an association between the family's income and what type of school would their children go to?
I'm not sure whether I'll be using a somer's d or chi-square.