Calculating and Interpreting Cronbach's Alpha Using SPSS

hello this is dr. grande welcome to my video on how to calculate chromebooks alpha also be showing you how to interpret chromebox alpha so let's start with a dataset in spss and chromebox alpha is a test of reliability of a scale so I have here 50 cases or 50 participants and 10 individual items that they responded to and these are in a Likert scale so they're 1 through 5 so 5 possible responses and these variables item 1 item 2 item 3 represent the actual responses given by the participant so on item one for participant 1 the response was 3 and I M 2 is 3 and so on chromebooks alpha is a popular method to determine reliability when working with Likert scales which is why I selected a Likert scale for this example so as you can see these are not scales all 10 items taken together would be a scale so items 1 through 10 represent a scale or sometimes called a sub scale of an instrument chromebox alpha does not work to determine the reliability of just one item you have to have a series of items in this case we have 10 now look at the variable view you can see these are loaded as ordinal and I know there's differing opinions regarding a Likert scale is it scale like this or is it ordinal I'm going to be a little more conservative for this demonstration and assume that Likert scale is ordinal but I'm aware that of course many people will believe that you can also use it as scale I believe that is true in certain circumstances let's just going to set that ordinal of course the case number is nominal so to analyze this to determine chromebox alpha you go to analyze then to scale then reliability analysis now I already have this populated so I'm going to return this back to default so this is what it looks like by default and the statistics nothing would be checked off this is how it look so what you want to have over here of course is your items so select all and then just subtract case and you can move them over one shot you can see you have all ten we leave it alpha looking for chromebox alpha and the statistics I like to run the descriptives with items scale and the scale item if deleted which is a very interesting output and then correlations continue and ok and let's take a look at the results you can see we have 50 cases and no missing values the chromebooks alpha is at point six zero six all right so that would be 0.7 is generally considered good 0.6 would be considered questionable so let's take a look at the different means you can see that the means and the standard deviations are all fairly close except for item 7 and item 10 they have a higher mean and higher standard deviation now we can see the interitum correlation matrix which is very straightforward and then we have the item total statistics and I want to draw attention in particular to the Alpha if an item is deleted so what this does is it tells you if you delete a particular item how does it affect the reliability of your scale now we know that it's point 6 right now so if you delete item 1 that would decrease it item to decrease it my fact all these items would decrease it except for item ten which actually increase it to point six nine eight which is almost at good and very close point six nine eight so what's it you know this would be as if we were building this instrument right so let's take the data we just are the results we just looked at from the output window and apply it so we'll analyze and go to scale reliability analysis now let's take out item ten now we kind of know what's going to happen right because it told us you can see it boosted chromebox alpha two to just about an acceptable level and now let's look and see if anything else can push it higher and of course now we see item seven stands out at point seven six which would definitely be a good chromebox alpha all these others of course lower items one through six and item items eight nine will give us a lower gum box alpha all right lower reliability so go back analyze again and this time we'll remove item seven we can see now chrome box office increased 2.76 using just a to the ten items that's important these are the ten items item ten nine seven notably missing and our chromebox alpha is 0.76 and you can see that item one being deleted now we'd move it up a little bit but and of course item nine a little bit but that's not that's not really something that is a huge increase so I think if you're looking for a good reliability coefficient the items to retain here would be one through six and eight nine and that's how we use to interpret chromebox alpha and that's how we use it to decide which items potentially to delete to increase the overall reliability of an instrument so as you can see chromebox alpha is a useful statistic to calculate but I do want to provide you with a few cautions about its use the first is that all other factors being held constant you'll get a higher chromebox alpha if your data is normally distributed then if it's positively or negatively skewed also if you have a high chromebox alpha for example say 0.95 that doesn't indicate that there is one dimension one construct - in order to determine how many constructs there are an instrument you'd have to use a factor analysis chromebox alpha is simply a measurement of internal consistency of reliability it is not an indicator of how many constructs are being measured by the scale I hope you found this video helpful as always if you have any questions or concerns feel free to contact me and I'll be happy to assist you

Channel: Dr. Todd Grande

Views: 302,862

Rating: 4.9213972 out of 5

Keywords: Cronbach's Alpha, SPSS (Software), alpha, reliability, likert scale, internal consistency, data, Statistics (Field Of Study), Grande

Id: Kz8OdR6lV44

Length: 8min 17sec (497 seconds)

Published: Mon Nov 10 2014