R predict() function returning too many values

Question

I am using the r predict function, and it is returning more values than I expected it too. I created a linear model for the data to predict MDC from PKWH, MDT, and MDT2, then I created new data for input values into the predict function. The original data for utility has 24 values for each column of MDC, PKWH, MDT, and MDT2.

    fit2 <- lm(MDC ~ MDT + MDT2 + PKWH*(1 + MDT + MDT2), data = 
    utility)
    predict <- predict(fit2, data = data.frame(PKWH = 9, MDT = 75, MDT2 
    = 5625))

I expected the predict() function to produce 1 predicted value for the inputs of PKWH = 9 | MDT = 75 | MDT2 = 5625, but it gave me these 24 values.

           1        2        3        4        5        6        7 
    56.67781 51.66653 45.05200 42.12583 38.98647 38.80904 42.60033 
           8        9       10       11       12       13       14 
    46.86545 49.51928 54.15163 61.54441 68.00122 49.17722 45.27917 
          15       16       17       18       19       20       21 
    42.88154 40.93468 38.39330 37.80963 39.47550 41.58780 42.94447 
          22       23       24 
    46.25884 49.27053 53.98732 

Also, when I plug the new input values to calculate the predicted value using the coefficients from the linear model, I get 55.42165 which is not found on the list of the 24 values from the predict() function.

Answer 1

first, I wouldn't name your result predict - you want to save that for the function. You need

predicted_data <- predict(fit2, newdata = data.frame(PKWH = 9, MDT = 75, MDT2 
= 5625))

It's not throwing an error because predict has a catch-all ( ... ) at the end where input to data is heading, but it's giving you the predictions for the data you fit the model with.