A | B |

SMALL DIFFERENCE BETWEEN PERCENTS | it is hard to attain significance with the chi square when this is small |

MCNEMAR | version of chi square appropriate for two-by-two repeated measures designs |

OBSERVED FREQUENCIES | it is easier to attain statistical significance when this differs greatly from theoretical frequencies |

EXPECTED FREQUENCIES | chi square should not be used when these are less than 5 per cell |

DF | determined by the rows and columns of the contingency table: (r - 1) * (c - 1) |

LARGE N | chi square can more easily attain statistical significance with this |

NON-PARAMETRIC TEST FOR SIGNIFICANCE (GOODNESS OF FIT) | chi square is this type of test |

NOMINAL | chi square reduces data to this level of measurement |

YATES | version of chi square corrected for use with small samples in two by two contingency table |

VALID | a measure which measures what it purports to measure |

RELIABLE | a measure which measures consistently |

SENSITIVE | a measure of a variable capable of detecting subjects who qualify |

SPECIFIC | a measure of a variable capable of excluding subjects who do not qualify |