On December 9, 1997, approximately 1600 Pacific standard time, a Cessna 150J, N51369, nosed over during a forced landing near Silverdale, Washington. The certified flight instructor and her student were not injured, but the aircraft, which was being leased by Alternate Air, Inc., sustained substantial damage. The 14 CFR Part 91 instructional flight, which departed Boeing Field, Seattle, Washington, about 45 minutes earlier, was being operated in visual meteorological conditions. No flight plan had been filed, and there was no report of an ELT activation. In a written statement, the flight instructor reported that during the recovery from the third of a series of power-off stalls, the engine began to run rough. The flight instructor stated that power was applied smoothly and carburetor heat was turned off as part of the stall recovery procedure. The engine continued to run rough and after approximately five seconds, the engine quit. The pilot stated that she took over the controls from the student pilot and attempted to restart the engine. During the first attempt, the propeller turned twice, but the engine did not fire. She then pulled the starter handle for a second restart attempt, but when she pulled it, the handle separated from the panel. At that point she picked out a field in which to land. The touchdown, which was made on the main gear, was uneventful, but when the nose gear was lowered onto the soft terrain, it sank in and the aircraft nosed over. During telephone interviews with both the flight instructor and student, they reported that carburetor heat was used during each power-off maneuver. A METAR observation taken at 1615, at Bremerton Airport, which is about seven miles south of the accident site, recorded a temperature of seven degrees Celsius and a dew-point of six degrees Celsius. According to the Carburetor Icing Probability Chart, the aircraft was operating in temperature and humidity conditions where severe icing is probably at both cruise and climb power. After the aircraft was recovered from the field, a Federal Aviation Administration Inspector from the Seattle Flight Standards District Office supervised an engine test run. The engine was started and run-up to 2500 rpm. The operation was smooth, with no discrepancies noted. At 1700 rpm, both magnetos were checked. The right hand magneto position indicated a 90 rpm drop, and the left hand magneto position indicated a 125 rpm drop. The carburetor heat control was positioned to the "on" position, and a 110 rpm drop was observed. The mixture control operated correctly. The engine pull-to-start knob was inspected and it was found that the knob had pulled off the metal shaft and the metal shaft had pulled away from the cable to the starter lever.