Skip to Main Content
An official website of the United States government
Epidemiology and Genomics Research Program

Percentage Energy from Fat Screener: Scoring Procedures

Recent Changes in Scoring

On July 19, 2004, an error was found in the data used to derive the scoring procedures. As a result, the scoring algorithm was changed and the values for the gender-specific portion sizes and the regression coefficients have been updated. You may download a new version of the SAS program used to calculate percentage energy from fat from the bottom of this page.

On October 12, 2003, the following changes were made to the calculations used to convert percentage of energy from fat from the screener data:

  1. Additional information about portion size was incorporated into the scoring algorithm.
  2. As before, a regression equation was used to estimate the individual's predicted percentage energy from fat intake. However, because of the additional portion size information, a new gender-specific set of regression coefficients was estimated.
  3. Whereas the regression coefficients presented in the earlier program were estimated using only individuals 50 to 69 years old, the new regressions were estimated using adults 18 and over.
  4. The data source for both portion size estimates and regression coefficients was the US Department of Agriculture's 1994-96 Continuing Survey of Food Intakes by Individuals.

If you need the previous version of the SAS program, please send an e-mail to rfab@mail.nih.gov.

Scoring Procedures

The following procedures were used to convert an individual's responses to an estimate of that individual's percentage energy from fat:

  1. The frequency reported categorically on the questionnaire was converted to the number of times consumed per day as shown below. In general, the midpoint of the frequency range was used.

    Frequency Response Times Per Day
    Never 0.0
    Less than once a month 0.018
    1-3 times per month 0.066
    1-2 times per week 0.214
    3-4 times per week 0.499
    5-6 times per week 0.784
    1 time per day 1.0
    2 or more times per day 2.0
  2. The age- and gender-specific portion sizes for each food (shown below) were multiplied by the frequency calculated in Step 1. These portion sizes were estimated from CSFII 94-96.

    Median Serving Sizes (g)

    Age Skim Milk Bacon Cold Cereal Fruit Mayo Dressing
    Males
    18-27 366.666667 25.000000 74.666667 131.750000 13.750000 36.720000
    28-37 250.000000 40.250000 61.500000 128.000000 13.750000 44.060000
    38-47 250.000000 32.000000 57.500000 123.200000 13.750000 31.250000
    48-57 245.000000 32.000000 56.000000 127.500000 13.750000 31.250000
    58-67 214.375000 27.000000 46.000000 122.000000 9.150000 29.400000
    68-77 198.937500 26.000000 39.000000 118.000000 13.750000 29.400000
    78+ 160.725000 24.000000 33.000000 114.250000 4.580000 29.380000
    Females
    18-27 245.000000 26.000000 50.000000 118.000000 13.750000 30.630000
    28-37 245.000000 25.000000 49.500000 118.000000 6.880000 29.400000
    38-47 244.800000 24.000000 44.000000 118.000000 9.170000 29.400000
    48-57 229.690000 24.000000 43.500000 118.000000 9.183333 29.400000
    58-67 196.000000 18.000000 33.000000 118.000000 6.110000 29.380000
    68-77 183.750000 19.500000 33.000000 112.427143 10.310000 29.380000
    78+ 183.750000 16.000000 33.500000 109.000000 4.580000 22.030000

    Median Serving Sizes (g)

    Age Eggs Fruit Juice Hot Dogs Cheese French Fries Margarine Rice
    Males
    18-27 92.000000 373.200000 114.000000 33.360000 112.500000 9.540000 213.625000
    28-37 92.000000 311.000000 85.500000 28.350000 114.000000 9.540000 195.000000
    38-47 92.000000 249.000000 88.000000 28.350000 100.000000 9.460000 166.000000
    48-57 92.000000 249.000000 114.000000 28.350000 100.000000 9.200000 165.000000
    58-67 92.000000 248.000000 57.000000 28.350000 85.500000 7.883333 165.000000
    68-77 80.000000 186.750000 57.000000 24.000000 85.500000 7.100000 158.000000
    78+ 80.000000 186.750000 57.000000 22.880000 97.000000 7.000000 158.000000
    Females
    18-27 80.000000 249.000000 57.000000 26.175000 79.500000 7.000000 158.000000
    28-37 80.000000 248.800000 57.000000 21.000000 70.000000 6.290000 158.000000
    38-47 69.000000 248.800000 57.000000 22.500000 70.000000 5.925000 158.000000
    48-57 80.000000 217.875000 114.000000 22.063333 70.000000 7.095000 155.000000
    58-67 68.000000 186.750000 57.000000 24.000000 66.000000 5.296667 122.250000
    68-77 56.000000 186.600000 57.000000 21.000000 70.000000 5.320000 158.000000
    78+ 46.000000 186.750000 57.000000 25.800000 64.000000 4.865000 83.000000
  3. The proportion of the margarine and butter added to foods that was regular fat was estimated. First, the frequency of the margarine and butter added was calculated (totfat in the table below). This is the sum of the frequencies for the three questions in question 1 for margarine and butter added to food (on bread, rolls, pancakes; vegetables including potatoes; and rice or pasta).

    totfat = sum(marg on bread, marg on veg, marg on rice)

    Then the information in the instrument's second question, "how often was reduced-fat margarine used?" was applied, using the following equivalents:

    Response to Question 2 Value of Regfat
    Didn't use or almost never totfat
    About 1/4 of the time totfat*.75
    About 1/2 of the time totfat*.50
    About 3/4 of the time totfat*.25
    Almost always or always 0
  4. The individual's percentage energy from fat was estimated by applying regression coefficients to each food item. The following equation was used to estimate percentage energy from fat (the dependent variable), for each gender:

    estpcalfat = intercept + (b1*cereal) + (b2*skim milk) + (b3*eggs) + (b4*bacon) + (b5*citrusjuice) + (b6*fruit) + (b7*hotdogs) + (b8*cheese) + (b9*friedpot) + (b10*mayo) + (b11*salad dressing) + (b12*rice) + (b13*reg.fat)

    The values used for each regression coefficient are in the following table:

    Estimated Regression Coefficients for Foods as Predictors of Percentage Energy from Fat, by Gender

    Parameter Male Female
    Intercept (b0) 30.795765 29.865870
    Cold Cereals (b1) -0.022086 -0.045171
    Skim Milk (b2) -0.009666 -0.010393
    Eggs (b3) 0.026997 0.036787
    Bacon or Sausage (b4) 0.109569 0.198808
    100% Fruit Juice (b5) -0.004946 -0.010141
    Fruit (b6) -0.009346 -0.012103
    Hot Dogs (b7) 0.040118 0.106686
    Cheese (b8) 0.069945 0.103239
    Fried Potatoes (b9) 0.024262 0.040374
    Regular Mayonnaise (b10) 0.145026 0.287044
    Regular Salad Dressing (b11) 0.114649 0.182758
    Rice (b12) -0.017017 -0.014224
    Margarine (b13) 0.167937 0.326702

Note that Question 3 was not used in this scoring. This last question by itself is correlated with percentage energy from fat, but it is unnecessary if the other questions are used to more precisely estimate the individual's percentage energy from fat.

SAS program to calculate Percentage Energy from Fat using the Short Screener.