The conclusions that can be drawn from the current study are limited in several ways. First, the data were collected to facilitate crisis intervention, rather than to answer specific research questions; therefore, the types of statistical analyses that could be applied to the data were limited. In addition, some call logs contained limited or incomplete data. For example, items may have gone unanswered if the caller declined to give particular pieces of information about him or herself or about the crisis situation. The data also consist of information from help-seeking individuals, who may differ from individuals who do not seek assistance in alleviating their crisis situations. For example, help-seeking individuals may have been more likely to disclose information about their problems or to have disclosed more serious issues such as physical abuse. The data were based on self-report and are thus potentially subject to biases such as social desirability or the selective underreporting of particular crisis issues. Underreported issues may have included experiences that involved some element of social stigma, such as having been sexually assaulted, or that involved the disclosure of criminal behavior perpetrated by the youth. Self-report biases may also have contributed to the relatively low rates of reported sexual abuse or assault among this sample as compared to reported rates of neglect and of physical, verbal and emotional abuse. Callers who wished to receive confidential help for their crisis situations may also have been reluctant to disclose issues that they believed would result in the contacting of law enforcement or protective service agencies, such as suicidality, violation of probation or parole, or parental abuse.
It is also difficult to determine whether the problematic issues of street youth, as recorded in the call log, occurred before or after the adolescent came to be away from the home. Therefore, the extent to which one can interpret a reported problem as a risk factor for becoming homeless, as opposed to a consequence of homelessness, is limited for this research sample. In addition, because the data are cross-sectional, the extent to which conclusions can be drawn about causality is limited.
Finally, the information provided by callers is subject to interpretation by the staff and trained volunteers who provide crisis intervention services. While NRS hotline staff and volunteer liners receive the same type and number of hours of initial training before taking calls and follow the same model of crisis intervention, the interpretation of data may vary depending on level of skill or amount of experience.