ELLEN GOLDBAUM'S APRIL 11 ARTICLE ON INTERNET ADDICTION WAS QUITE INTERESTING. HOWEVER, IN SPITE OF THE HIGHLY PROBABLE EXISTENCE OF ADDICTION, AS INDICATED BY THE CITED STUDY, THE EXTENT OF SUCH A PROBLEM (17 PERCENT OF RESPONDENTS SPENDING OVER 40 HOURS PER WEEK ONLINE) CANNOT BE TAKEN SERIOUSLY. IN ALL FAIRNESS, IT MUST BE COMMENDED THAT THE AUTHORS OF THE STUDY RECOGNIZED THAT IT WAS PLAGUED BY SELECTION BIAS. YET, THE ARTICLE WAS CONTENT TO LEAVE THE READER WITH THE IMPRESSION THAT THIS IS NEVERTHELESS A WIDESPREAD PROBLEM.

THE TRUTH IS THAT THE PRESENCE OF SELECTION BIAS PERMITS ONLY THE CONCLUSION THAT INTERNET ADDICTION EXISTS. FIRST, CONSIDER THAT ONLY A FRACTION OF THE USERS OF THE NET WOULD COME UPON THE STUDY SITE. SECOND, OF THOSE THAT DID DISCOVER THE SITE, A FRACTION WOULD IGNORE IT COMPLETELY, EITHER FOR REASONS OF TIME OR INTEREST. FINALLY, OF THOSE TAKING NOTE OF THE STUDY ITSELF, ONLY A FRACTION WOULD BOTHER TO RESPOND TO THE QUESTIONNAIRE. THIS IS EXACTLY THE PROBLEM: ADDICTED USERS WOULD BE THE MOST LIKELY TO FIND AND RESPOND TO THE SURVEY. MOST OTHERS ARE IN SEARCH OF SOMETHING SPECIFIC, AND COULDN'T BE BOTHERED.

AMONG THE MANY PARAMETERS WHICH MUST BE ESTIMATED ARE THE FRACTION OF USERS WHO ACTUALLY RESPOND. THE TRUTH IS THAT THE MARGIN OF ERROR IS SO GREAT IN ESTIMATING SUCH THINGS THAT NOTHING OF SUBSTANCE CAN BE CONCLUDED REGARDING THE EXTENT OF THE PROBLEM. THIS SHOULD HAVE BEEN MORE CLEARLY EMPHASIZED IN THE ARTICLE.

SINCERELY,

STEPHEN R. BESCH, PH.D.