# May I ask for some guidance?

#### KurogamiMasaru

##### New Member
I just need to verify if what I am doing is right. I conducted a survey regarding how much the respondents agree on the effectiveness of a certain program using three components.

I wanted to see if there is a significant number of the respondents who answered at least "agree" from the scale.

I used one-sample chi-square from SPSS (I am not sure but apparently it uses Chi-square for goodness of fit.)

My understanding is that if the null hypothesis is rejected, then it means that the trend of the responses are actually geared towards "agree" and not evenly distributed among the other responses, and that the trend is significant.

Did I follow a good path? I only have basic understanding of statistics for research.

#### Karabiner

##### TS Contributor
Did you have 2 categories, "at least agree" versus "did not agree"?

By default, the SPSS one sample Chi² test assumes that you want to know whether
there is a significant deviation from a 50% versus 50% distribution. If you have a
significant result, look at the descriptive statistics, which of the two proportions is
the higher one.

#### Dason

Please provide some information on what the options available for the question were

#### KurogamiMasaru

##### New Member
Did you have 2 categories, "at least agree" versus "did not agree"?

By default, the SPSS one sample Chi² test assumes that you want to know whether
there is a significant deviation from a 50% versus 50% distribution. If you have a
significant result, look at the descriptive statistics, which of the two proportions is
the higher one.
thank you for responding,

so does this test only work for dichotomous variables and not for scales?

#### KurogamiMasaru

##### New Member
Please provide some information on what the options available for the question were
thank you for the response,

right,
sorry for missing to provide that,
it is a scale of 1-5
1 = strongly disagree, 5 = strongly agree,

agree is at 4 and that is the minimum requirement for the objective to be met,
there was a relatively high number of responsents who answered 4 but I wanted to see if it was a significant number of respondents,

#### Dason

Rejecting the null on the test you were doing would only tell you that at least one of the responses had a different proportion.

#### KurogamiMasaru

##### New Member
Rejecting the null on the test you were doing would only tell you that at least one of the responses had a different proportion.
and then base it on the descriptive to specify which one and the proportion is not actually significant?

#### Karabiner

##### TS Contributor
thank you for responding,

so does this test only work for dichotomous variables and not for scales?
You said "I wanted to see if there is a significant number of the respondents who answered at least "agree" from the scale.".
I supposed that you therefore collapse all responses which at least "agree".
It would be useful if you specify what you really want to find out here.