Presentation is loading. Please wait.

Presentation is loading. Please wait.

A. Chi-square (Goodness of fit) Question answered: Is the actual distribution of items into categories different from what you could get by chance?

Similar presentations


Presentation on theme: "A. Chi-square (Goodness of fit) Question answered: Is the actual distribution of items into categories different from what you could get by chance?"— Presentation transcript:

1 a. Chi-square (Goodness of fit) Question answered: Is the actual distribution of items into categories different from what you could get by chance?

2 a. Chi-square (Goodness of fit) Chi-square can be used to answer question such as: Do people match the colors yellow, red, and blue to the vowels [i] [ ɑ ] [u] in a particular pattern (e.g. [ ɑ ] is red)? In what register (spoken, academic, or newspaper) is turn the Noun on versus turn on the Noun more common?

3 a. Chi-square (Goodness of fit) Chi-square compares the observed counts against the expected counts.

4 a. Chi-square (Goodness of fit) (Invented) frequency of use of dude in four million word spoken corpora: US NZ AU UK 15 9 11 5 Random distribution would be: Observed (what the actually did) US NZ AU UK 10 10 Expected (what you would expect by random chance) 15 9510 USNZ AU UK

5 a. Chi-square (Goodness of fit) chi-square = 2.77 degrees of freedom = 3 probability = 0.429 We want this to be large We want this to be small Degrees of freedom is the number of rows containing the observed responses minus one (1-1=0) plus the number of columns that contain observed responses minus one (4-1=3), or in other words, 3. http://www.vassarstats.net/csfit.html

6 a. Chi-square (Goodness of fit) For example, multiple choice questions. Jill loves the taste of coffee, if she should she’d c_ffify everything. ObservedExpected A-c[æ]ffify186 123 B-c[^]ffify113 123 C-c[a]ffify70123 Is 186, 113, 70 really different from what random choice would give?

7 a. Chi-square (Goodness of fit) Results: chi-square = 55.919 degrees of freedom = 2 probability = 0.0005 How do you calculate the degrees of freedom?

8 a. Chi-square (Goodness of fit) Remember: Chi-square can only be used with counts, not percentages, or non-whole numbers.

9 a. Chi-square (Goodness of fit) Practice: Is the word glistening used more often in one register (as shown in COCA) than another? SECTIONSPOKENFICTIONMAGAZINENEWSPAPERACADEMIC PER MIL0.412.02.82.10.6 SIZE (MW)76.669.678.173.473.0 FREQ3283321915643 What is the problem with these data? How can you correct it?

10 a. Chi-square (Goodness of fit) Practice: Is the word glistening used more often in one register (as shown in COCA) than another? http://www.physics.csbsju.edu/stats/contingency_NROW_NCOLUMN_form.htm l SECTIONSPOKENFICTIONMAGAZINENEWSPAPERACADEMIC PER MIL0.412.02.82.10.6 SIZE (MW)76.669.678.173.473.0 FREQ3283321915643 What is the problem with these data? How can you correct it?

11 a. Chi-square (Goodness of fit) Are women overrepresented in foreign language classes? Total 4414 http://www.vassarstats.net/csfit.html ObservedExpected Men24132207 Women20012207

12 a. Chi-square (Goodness of fit) What if there were 55% women and 45% men in university? Total 4414 Men: 4414 x.45 = 1986 Women: 4415 x.55 = 2428 http://www.physics.csbsju.edu/stats/contingency_NROW_NCOLUMN_f orm.html ObservedExpected Men24131986 Women20012428

13 Chi-square Goodness of fit Chi-square Has one independent variable Asks if observed differs from expected Test of independence Chi-square Has two or more variables Asks if the variables are dependent on each other or are they independent?

14 Chi-square (Test of independence) Is snuck (vs. sneaked) and Americanism? Research Question: Is the use of sneak and snuck independent from dialect (US vs. UK)?

15 a. Chi-square (Goodness of fit) Tells us whether something happened more often than chance would predict http://www-user.uni- bremen.de/~anatol/qnt/qnt_chi.html Use with multiple choice questions, percentage of time respondents choose specific choice, more corpora or frequency data

16 Chi-square exercise (in class)


Download ppt "A. Chi-square (Goodness of fit) Question answered: Is the actual distribution of items into categories different from what you could get by chance?"

Similar presentations


Ads by Google